osgeo4 backup is failing

[robe]

This might be result of upgrade or something.

Error: Failed to run: zfs destroy osgeo7/containers/secure@snapshot-for-osgeo4: cannot destroy snapshot osgeo7/containers/secure@snapshot-for-osgeo4: dataset is busy
Error: Failed to run: zfs destroy osgeo7/containers/wordpress@snapshot-for-osgeo4: cannot destroy snapshot osgeo7/containers/wordpress@snapshot-for-osgeo4: dataset is busy
Error: Failed to run: zfs destroy osgeo7/containers/dronie-server@snapshot-for-osgeo4: cannot destroy snapshot osgeo7/containers/dronie-server@snapshot-for-osgeo4: dataset is busy

It sometimes resolves itself but sometimes doesn't and then a reboot of osgeo7 is required or explicitly unmount / mount the containers.

I think the script also needs some work as it looks like it deleted the backups even though taking a snapshot failed.

[robe]

Okay this is a different issue than what I though. Usually when this kind of thing happens with making snapshots, I can't make snapshots at all and it gives a different error.

I can snapshot these servers fine with

lxc snapshot secure
lxc snapshot wordpress
lxc snapshot dronie-server

However if I try to delete

lxc rm dronie-server/for-osgeo4

I get this error:

 Failed to run: zfs destroy osgeo7/containers/dronie-server@snapshot-for-osgeo4: cannot destroy snapshot osgeo7/containers/dronie-server@snapshot-for-osgeo4: dataset is busy

This I have only seen happen if osgeo4 is in the middle backing up the container in question.

Checking osgeo4 - it is in middle of backup of pretalx on osgeo3 which shouldn't impact osgeo7

 sudo ps -faux | grep "lxc copy"

shows:

lxc copy osgeo3:pretalx/for-osgeo4 pretalx-backup

So only thought I have is osgeo4 must still have a hold on it when it tried to delete the snapshot to reuse the name. I suspect rebooting osgeo4 should resolve this. But should wait till it's done with backups.

[robe]

closing this out since last scheduled backup on these ran fine. dronie-server one I manually ran since that only backs up every 2 days.